Attribution is the business of deciding which marketing touch deserves credit for a sale. In categories with short, online, single-person purchases it is a solvable problem. Travel is none of those things, which is why travel attribution generates more confident wrong answers than almost any other area of marketing.
Why travel is unusually difficult
Four characteristics compound:
- Long consideration. Weeks to months between first interest and payment, often crossing cookie lifetimes.
- Multiple devices. Discovered on a phone during a commute, researched on a laptop at work, discussed on a tablet at home.
- Multiple people. One person researches, another pays, a third has veto power over the dates.
- Offline closing. The booking is confirmed on a call or a WhatsApp thread, not on a checkout page.
Any single one of these would degrade attribution. Together they mean a default analytics setup will misattribute a large share of travel revenue, usually in the same predictable direction.
What the common models actually do
It helps to be precise about what each model claims.
| Model | Credits | In travel, this |
|---|---|---|
| Last click | Final touch | Over-credits brand search and retargeting |
| First click | Initial touch | Over-credits discovery channels like Meta and YouTube |
| Linear | All touches equally | Treats a three-second video view as equal to a package page visit |
| Time decay | Recent touches more | Reasonable default for long cycles |
| Data-driven | Modelled contribution | Best available, needs volume and clean data |
The most common failure is not choosing the wrong model. It is comparing numbers produced under different models — Google Ads reporting its own conversions on a data-driven model, Meta reporting on its view-through window, GA4 reporting on something else again — and being surprised that they disagree.
A practical approach
Rather than searching for a perfect model, run three layers that answer different questions.
Layer 1 — The CRM is the source of truth for revenue
Your CRM knows which enquiries became bookings and what they were worth. That is not in dispute. Capture the source and campaign at the point of enquiry, store it permanently on the lead record, and report booked revenue by source from there. This number is defensible in a way no platform number is.
Layer 2 — Platform data optimises delivery
Google and Meta conversion data exists primarily to steer their algorithms, not to settle internal debates about credit. Feed each platform the outcomes it needs — qualified leads and bookings via offline import — and then let each optimise against its own view.
Do not add platform-reported conversions together. The overlap is real and often substantial.
Layer 3 — Ask people, systematically
A single "How did you first hear about us?" field, asked during the sales conversation and recorded as a structured field, is one of the most valuable and least used attribution instruments in travel.
It is imperfect — people misremember, and they under-report advertising — but it captures things no tracking can: a friend's recommendation, a hoarding, a travel fair, a mention in a WhatsApp family group. Over a few hundred enquiries the pattern is genuinely informative.
The test that settles arguments
When attribution disputes stall — typically over whether a demand-generation channel deserves credit it cannot prove — stop modelling and run an incrementality test.
Turn the channel off in one region or for a defined period, keep everything else stable, and observe what happens to total enquiries and bookings, not just that channel's reported numbers. If total volume holds, the channel was taking credit for demand that existed anyway. If it falls, the channel was creating demand that attribution was failing to see.
It is blunt, it costs something to run, and it answers the question definitively — which no amount of model-switching will.
What good looks like
A travel business with attribution under control can state, for any month: total spend by channel, enquiries by source, qualified enquiries by source, bookings and booked revenue by source, and cost per booking by source — with the honest caveat that a proportion is unattributable.
That last clause matters. Attribution is a decision-making tool, not an accounting system. It needs to be accurate enough to allocate the next rupee correctly. It does not need to be perfect, and pretending it is perfect is how businesses defund channels that were quietly working.
Want this applied to your own funnel?
The Free Travel Growth Audit reviews your ads, landing pages, tracking and lead follow-up, and tells you which of these is currently costing you the most.
Book a Free Travel Growth Audit